-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathfeature_functions.py
130 lines (103 loc) · 3.93 KB
/
feature_functions.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
# -*- coding: utf-8 -*-
from __future__ import absolute_import, division, print_function
import numpy as np
import pandas as pd
from geopy.distance import vincenty
from geopy.geocoders import Nominatim
u''' 特征工程部分 函数'''
dtypes_train = {
0: 'uint32',
1: 'uint32',
2: 'float32',
3: 'float32',
4: 'float32',
5: 'uint8',
6: 'float32',
7: 'float32',
8: 'uint8',
9: 'float32'
}
dtypes_test = {
0: 'uint32',
1: 'uint32',
2: 'float32',
3: 'float32',
4: 'float32',
5: 'uint8',
6: 'float32',
7: 'float32',
8: 'uint8'
}
def base_feature(item_np,feature_name):
v = {}
v[feature_name + '_mean'] = item_np.mean()
v[feature_name + '_std'] = item_np.std()
v[feature_name + '_skew'] = item_np.skew()
v[feature_name + '_kurt'] = item_np.kurt()
v[feature_name + '_min'] = item_np.min()
v[feature_name + '_max'] = item_np.max()
v[feature_name + '_median'] = item_np.median()
v[feature_name + '_q25'] = item_np.quantile(0.25)
v[feature_name + '_q75'] = item_np.quantile(0.75)
v[feature_name + '_var'] = item_np.var()
return v
def base_feature_extract(df,columns):
ret = {}
for col in columns:
i_ret =base_feature(df[col],col)
ret.update(i_ret)
return ret
geolocator = Nominatim()
def dtype_columns(dtype_dict):
ret = {}
for i,v in dtype_dict.items():
c = ret.get(v,[])
c.append(i)
ret[v] = c
return ret
# dtypes_train = dtype_columns(dtypes_train)
# dtypes_test =dtype_columns(dtypes_test)
def tranfrom_itemgeter(item,dtypes_dict,sorted_list=[1,2]):
ret = item.astype(dtypes_dict,copy=True)
# for type,columns in dtypes_dict.items():
# item[columns] = item[columns].astype(type)
ret.sort_values(sorted_list,inplace=True)
return ret
def callstate(item):
'''通话记录'''
callstate_init = {0:0,1:0,2:0,3:0,4:0}
uniq,count = np.unique(item[8],return_counts=True)
sumcount = count/(sum(count)+0.0)
s = dict(zip(uniq,sumcount))
callstate_init.update(s)
ret = {'callstate_{0}'.format(k):v for k,v in callstate_init.items()}
return ret
def geo_info(item):
'''时间间隔'''
time_gap =(item[1:][1] - item.shift(1)[1:][1])
end_local = item[1:][[3,4,7]]
start_local = item.shift(1)[1:][[3,4,7]]
start_local.columns = ['start_longitude','start_latitude','start_speed']
end_local.columns = ['end_longitude','end_latitude','end_speed']
loacl_se =pd.concat([start_local,end_local,time_gap],axis=1)
loacl_se['distance_se'] = loacl_se.apply(lambda x: vincenty((x['start_latitude'],x['start_longitude']),(x['end_latitude'],x['end_longitude'])).m,axis=1)
loacl_se['distance_speed'] = loacl_se['distance_se']/loacl_se[1]*3600/1000
# 平均加速度
loacl_se['avg_acc'] = (loacl_se['start_speed'] - loacl_se['end_speed'])/loacl_se[1]
loacl_se['sub_start_acc'] = (loacl_se['start_speed'] - loacl_se['distance_speed'])/loacl_se[1]/2
loacl_se['sub_end_acc'] = (loacl_se['end_speed'] - loacl_se['distance_speed']) / loacl_se[1]/ 2
# 绝对值
loacl_se['avg_acc_l2'] = np.sqrt(np.abs((loacl_se['start_speed']**2 - loacl_se['end_speed']**2)/loacl_se[1]))
loacl_se['sub_start_acc_l2'] = np.sqrt(np.abs(loacl_se['start_speed']**2 - loacl_se['distance_speed']**2)/loacl_se[1]/2)
loacl_se['sub_end_acc_l2'] = np.sqrt(np.abs(loacl_se['end_speed'] - loacl_se['distance_speed']) / loacl_se[1]/ 2)
# 时间超过60s
# 二阶特征
# loacl_se['sub_time_ada'] = (loacl_se[1] > 60).cumsum()
# 构造一阶特征
# 速度类
speed_columns = ['distance_speed','avg_acc','sub_start_acc','sub_end_acc','avg_acc_l2', 'sub_start_acc_l2','sub_end_acc_l2']
res = base_feature_extract(loacl_se,speed_columns)
return res
'''要处理的函数 注册进 feature_func'''
'''注册方法 {'方法名字'：方法的统计特征} '''
feature_func = {u'callstate_stats': callstate,'geo_speed':geo_info}